Publish in Scimago's Q1 and Q2 Journals

Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4196

Browse

Now showing 1 - 20 of 215

Open Access
A Comparative Investigation of Infiltration and Channel Roughness of Ephemeral and Perennial Streams in a Mountainous Catchment
(John Wiley, 2025-06) Khaniya, B; Gomes, P.I.A; Perera,M. D.D; Wai, O, W.H
Infiltration and channel roughness, two major factors that govern stream discharge, were studied in similar-sized ephemeral and perennial streams in a mountainous tropical catchment. Seasons were defined based on two ephemeral flow conditions, i.e., with (wet season) and without (dry season) surface flow. A stream was divided transversely into low-flow areas (close to the thalweg) and high-flow areas (close to the channel margin). The highest average infiltration (~50 mm/h) was observed in the low flow areas around the thalweg of ephemeral streams in the dry season and was significantly higher than for any other spatial scale or temporal period. The infiltration in high-flow areas did not show a statistically significant difference between the two stream types, and surprisingly, perennial streams in the dry season showed higher infiltration than ephemeral streams. Since sediment moisture and organic content showed negative and positive correlations with infiltration, respectively, for both stream types and ephemeral streams showed statistically significant negative correlations between litter and infiltration during the dry season, the low infiltration in ephemeral high flow areas was attributed to stream-type dependent litter processing. The litter of ephemeral stream high-flow areas was subject to partial decomposition due to rapid drying and had residue of previously buried litter. Ephemeral channels were two to three times rougher than perennial channels. Standing crop biomass and mean particle size increased stream roughness in both stream types but were less prominent in ephemeral streams due to the presence of litter. The study demonstrated that litter has a special role in defining the infiltration pattern, channel roughness, and flood control potential of ephemeral streams.
Open Access
A Context-Aware Doorway Alignment and Depth Estimation Algorithm for Assistive Wheelchairs
(Multidisciplinary Digital Publishing Institute (MDPI), 2025-07-17) Tennekoon, S; Wedasingha, N; Welhenge, A; Abhayasinghe, N; Murray, I
Navigating through doorways remains a daily challenge for wheelchair users, often leading to frustration, collisions, or dependence on assistance. These challenges highlight a pressing need for intelligent doorway detection algorithm for assistive wheelchairs that go beyond traditional object detection. This study presents the algorithmic development of a lightweight, vision-based doorway detection and alignment module with contextual awareness. It integrates channel and spatial attention, semantic feature fusion, unsupervised depth estimation, and doorway alignment that offers real-time navigational guidance to the wheelchairs control system. The model achieved a mean average precision of 95.8% and a F1 score of 93%, while maintaining low computational demands suitable for future deployment on embedded systems. By eliminating the need for depth sensors and enabling contextual awareness, this study offers a robust solution to improve indoor mobility and deliver actionable feedback to support safe and independent doorway traversal for wheelchair users.
Open Access
A cost effective machine learning based network intrusion detection system using Raspberry Pi for real time analysis
(PLOS ONE, 2025-12-29) Wijethilaka R.W.K.S; Yapa, K; Siriwardena, D
In an increasingly interconnected world, the security of sensitive data and critical operations is paramount. This study presents the development of a Network Intrusion Detection System (NIDS) that analyzes both inbound and outbound network traffic to detect and classify various cyber attacks. The research begins with an extensive review of existing intrusion detection techniques, highlighting the limitations of traditional methods when addressing the unique security challenges posed by distributed networks. To overcome these limitations, advanced machine learning algorithms, including Random Forest, Long Short Term Memory (LSTM) networks, Artificial Neural Networks (ANN), XGBoost, and Naive Bayes, are employed to create a robust and adaptive intrusion detection system. The practical implementation utilizes a Raspberry Pi as the central processing unit for real time traffic analysis, supported by hardware components such as Ethernet cables, LEDs, and buzzers for continuous monitoring and immediate threat response. A comprehensive alert system is developed, sending email notifications to administrators and activating physical indicators to signify detected threats. Our proposed NIDS achieves 96.5 detection accuracy on the NF-UQ-NIDS dataset, with a significantly reduced false positive rate after applying SMOTE. The system processes real time network traffic with an average response time of 50 milliseconds, outperforming traditional IDS solutions in accuracy and efficiency. Evaluation using the NF-UQ-NIDS dataset demonstrates a significant improvement in detection accuracy and response time, establishing the system as an effective tool for safeguarding networks against emerging cyber threats.
Open Access
A cross-category analysis of high impact low occurrence (HILO) disasters
(Elsevier Ltd, 2026-03-19) Samaraweera, U; Kulatunga, U; Dias, P
This paper explores six High Impact Low Occurrence (HILO) disasters, generating insights from five different categories associated with them, namely causes (geophysical, technological, biological, sociological), phases (preparedness, response, recovery), dimensions (socio-economics, governance, equity), sectors (health, education, infrastructure, economy) and national contexts with differing levels of economic development. The process involved the generation of a questionnaire, based on a literature review; and the subsequent analysis and discussion of the questionnaire responses made by six experts nominated by six academies of science in Asia. The findings highlight the limitations of probabilistic, frequency-based risk models for HILO disasters and instead emphasise the importance of scenario-based (worst-case) analyses; mechanisms that preserve inter‐generational knowledge, institutional continuity and community‐based early‐response networks; strengthening community resilience while ensuring equity; and making appropriate investments for increasing preparedness, if not through structural interventions, at least through sustained awareness programs and periodic drills. Theoretical contributions include arguments that institutional capacity, governance quality, and social resilience are more decisive determinants of HILO event outcomes than probabilistic risk analyses; and that effective preparedness depends more on anticipatory planning, scenario-based training and institutionalised memory rather than experiential learning; thus advancing HILO theory beyond event-centred and frequency-driven interpretations towards a more governance- and resilience-oriented understanding.
Open Access
A novel application with explainable machine learning (SHAP and LIME) to predict soil N, P, and K nutrient content in cabbage cultivation
(Elsevier B.V., 2025-03-06) Abekoon, T; Sajindra, H; Rathnayake, N; Ekanayake, I, U; Jayakody, A; Rathnayake, U
Cabbage (Brassica oleracea var. capitata) is commonly cultivated in high altitudes and features dense, tightly packed leaves. The Green Coronet variety is well-known for its robust growth and culinary versatility. Maximizing yield is crucial for food sustainability. It is essential to predict the soil’s major nutrients (nitrogen, phosphorus, and potassium) to maximize the yield. Artificial intelligence is widely used for non-linear predictions with explainability. This research assessed the predictive capabilities of soil nitrogen, phosphorus, and potassium levels with explainable machine learning methods over an 85-day cabbage growth period. Experiments were conducted on cabbage plants grown in central hills of Sri Lanka. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) were used to clarify the model’s predictions. SHAP analysis showed that high feature values of the number of days and plant average leaf area negatively impacted for nutrient predictions, while high feature values of leaf count and plant height had a positive effect on the nutrient predictions. To validate the results, 15 greenhouse-grown cabbage plants at various growth stages were selected. The nitrogen, phosphorus, and potassium levels were measured and compared with the predicted values. These insights help refine predictive models and optimize agricultural practices. A user-friendly application was developed to improve the accessibility and interpretation of predictions. This tool is a user-friendly platform for end-users, enabling effective use of the model’s predictive capabilities.
Open Access
A physics-informed machine learning for detecting suspicious satellite maneuvers (orbital manipulation)
(Elsevier B.V., 2026) Karunathilake K.K.H; Abeywardena, K. Y ; Vecchini, S
Satellite systems have become prime targets for cyberthreats given their critical role in global infrastructure and general lack of security. Among these, orbital manipulation, a form of satellite hijacking, is a particularly severe threat that can disrupt essential operations and impact national security. To address these concerns, this research proposes an Artificial Intelligence (AI)-based anomaly detection system that utilizes Machine Learning (ML) models to analyze telemetry data for possible orbital manipulations with a multi-gate physics architecture grounded in orbital mechanics, to verify that detected anomalies are kinematically inconsistent and are therefore genuine integrity failures. This research demonstrates that temporal-based models like LSTM are essential for this domain, achieving high recall rates which are then validated by the physics component. While the framework includes multiple physical constraints, this study specifically validates the energy-based Vis-Viva gate, with the Tsiolkovsky and Angular Momentum gates established as architectural designs for future verification. This study concludes that successful AI deployment in orbital cybersecurity requires a comprehensive approach that integrates domain-specific context and physics-informed validation beyond traditional performance metrics
Open Access
A physics-informed machine learning for detecting suspicious satellite maneuvers (orbital manipulation)
(Elsevier B.V., 2026-07) Karunathilake K.K.H.; Abeywardena, K. Y; Vecchini, S
Satellite systems have become prime targets for cyberthreats given their critical role in global infrastructure and general lack of security. Among these, orbital manipulation, a form of satellite hijacking, is a particularly severe threat that can disrupt essential operations and impact national security. To address these concerns, this research proposes an Artificial Intelligence (AI)-based anomaly detection system that utilizes Machine Learning (ML) models to analyze telemetry data for possible orbital manipulations with a multi-gate physics architecture grounded in orbital mechanics, to verify that detected anomalies are kinematically inconsistent and are therefore genuine integrity failures. This research demonstrates that temporal-based models like LSTM are essential for this domain, achieving high recall rates which are then validated by the physics component. While the framework includes multiple physical constraints, this study specifically validates the energy-based Vis-Viva gate, with the Tsiolkovsky and Angular Momentum gates established as architectural designs for future verification. This study concludes that successful AI deployment in orbital cybersecurity requires a comprehensive approach that integrates domain-specific context and physics-informed validation beyond traditional performance metrics.
Embargo
Accessibility and usability of virtual learning platforms: Lived experiences of visually impaired undergraduates in Sri Lanka
(Elsevier Ltd, 2026-03-12) Rajapakshe, W; Wickramaarachchi, C; Alwis, M.K. S.S; Amarasinghe, A.A. M.L; Jayasekara, P.N; Jayasekara, P.T
This study explores the accessibility and usability of virtual learning platforms of visually impaired undergraduate students in Sri Lanka, focusing on their lived experiences, use of assistive technologies, and institutional support mechanisms. As online learning becomes increasingly prevalent, accessibility and inclusive challenges persist, particularly in developing countries with limited infrastructure and institutional support. Despite the availability of assistive technologies, visually impaired learners frequently encounter barriers, including poorly designed platforms, limited usability of screen readers, and inadequate institutional guidance. Addressing a critical research gap, this study investigates how visually impaired undergraduates experience and navigates virtual learning environments to identify accessibility barriers, enabling practices, and context-specific strategies for inclusive digital learning. Using a qualitative approach, semi-structured interviews were conducted with fifteen visually impaired university students across Sri Lanka. Thematic analysis revealed five core themes: barriers and challenges to effective virtual learning, preferred virtual platforms, accessibility features and tools, facilitators of learning success, and strategies to optimise the learning environment. These findings illuminate how systemic inequalities, infrastructural limitations, and institutional neglect collectively constrain the digital learning experience for visually impaired students, while also highlighting enabling practices that foster access and inclusion. The study's originality lies in foregrounding student-led insights in a developing country context and integrating practical, context-specific recommendations for platform developers, educators, and policymakers. By centering the voices of visually impaired learners, this research contributes unique and actionable knowledge to the field of inclusive digital education.
Open Access
Achieving zero hunger: A global policy lens on food security drivers and income group disparities
(Elsevier B.V., 2026-01-19) Pulle, N; Sampath, P; Perera, S; Wijayaweera, D; Jayathilaka, R
Many countries struggle to meet their daily dietary requirements despite numerous attempts to address the existing demand. Consequently, this study collectively analyses the impact of urbanisation, renewable energy, greenhouse gas emissions, population growth, gross domestic product per capita and agricultural land on food production relying on Sen’s Entitlement Theory, thus providing insights to resolve the long-standing issue of food insecurity, and support the achievement of the Sustainable Development Goals. The study utilises a stepwise panel ordered Probit model on 146 countries, for the years 1993 to 2023. It further categorises the food production index into three categories of food security as; low, moderate and high, thereby enabling discussion of the likelihood of a country falling into one of the aforementioned food security categories over the years. Urbanisation, agricultural land, and the dummy variables introduced to represent the income groups have been identified to have a significant and favourable relationship with the food production index. In contrast, the greenhouse gas emissions and renewable energy variables have a significantly inverse impact on the food production index. This makes a unique contribution to the existing body of literature, especially by comparing odds over the years, across different food secure categories, countries, and their specific income levels. This study enables policymakers to gain a comprehensive historical perspective on each case. This study further promotes the Sustainable Development Goals, highlighting areas where these goals have been negatively impacted. Additionally, the study discusses optimised investment allocations, agricultural research and development, agricultural technology, climate resilient farming, and sustainable urbanisation planning as solutions for extreme cases
Embargo
Addressing Child Labour in SAARC: The Synergy of Education, Health and Economic Growth Towards SDGs
(John Wiley and Sons, 2025-11-09) Muthugala, H; Magammana, T; Perera, A; Bandara, A; Jayathilaka, R
Child labour remains a critical socio-economic challenge in the SAARC region, closely linked to sustainable development goals (SDGs). This study investigates the determinants of child labour by examining the roles of education, health and economic growth using a robust methodological framework. The analysis captures the non-linear country-specific relationships between these variables and child labour, employing advanced methodological approaches, including multiple polynomials, stepwise and simple polynomial regression. The findings reveal a complex interplay of factors, with each variable showing positive and negative effects on child labour in country-specific contexts. Improved access to education generally reduces child labour, but disparities in quality and affordability can have the opposite effect. Health improvements significantly lower child labour rates, yet unequal healthcare access perpetuates exploitation among vulnerable groups. Economic growth shows dual effects: it promotes adult employment and alleviates poverty, yet unregulated expansion in specific sectors can heighten the demand for child labour. This study makes a novel contribution by integrating socio-economic determinants with child labour within a regional framework, providing actionable insights while aligning with SDGs 3, 4, 8 and 8.7. Key policy recommendations include fostering regional collaboration, ensuring access to free education, enacting and enforcing new laws, improving healthcare infrastructure and promoting inclusive and sustainable economic growth. These measures align with global SDG commitments but aim to secure a brighter future for the region's children by achieving these goals by 2030.
Open Access
Advancing Object Detection: A Narrative Review of Evolving Techniques and Their Navigation Applications
(Institute of Electrical and Electronics Engineers Inc., 2025-03-17) Tennekoon, S; Wedasingha, N; Welhenge, A; Abhayasinghe, N; Murray Am, I
Object detection plays a pivotal role in advancing computer vision systems by enabling machines to perceive and interact intelligently with their environments. Despite significant advancements, comprehensive exploration of its evolution and applications in navigation remains underrepresented. This review paper examines the evolution of object detection technologies, from early methodologies to contemporary advancements, and their critical role in navigation tasks. The emphasis was on the significance of contextual learning in enhancing object detection performance by leveraging spatial and temporal information. Furthermore, the limitations of conventional approaches that rely heavily on hand-engineered features are examined. It is then demonstrated that contextual learning facilitates automated feature extraction, resulting in improved accuracy exceeding a 50% increase and adaptability in diverse applications. The review concludes by outlining future trends and opportunities for further advancements in object detection and, underscoring its transformative impact on autonomous navigation and beyond. In summary, this review contributes to a comprehensive understanding of object detection technologies by offering insights into their evolution, highlighting their applications in navigation, and providing guidance for future research in context-aware systems.
Open Access
AI-Driven Adaptive Security for Sensor Networks: Next-Generation Firewalls for Attack Detection
(John Wiley and Sons, 2025-07-25) Meegammana, N.W; Fernando, H
Sensor networks are increasingly critical in modern smart environments; however, their limited computational resources expose them to sophisticated cyber threats. Traditional static firewalls and computationally intensive deep learning models are impractical for securing such networks. This study proposes an adaptive next-generation firewall (NGFW) that dynamically switches between shallow and deep AI models based on real-time network load and resource availability. Four neural network models were trained using 20 and 40-feature subsets of the UNSW-NB15 dataset. Two runtime strategies (i) on-demand model loading and (ii) preloaded model switching were developed and evaluated through simulation under real-time conditions. Experimental results indicate that the preloaded approach achieves up to 96% accuracy, 98% precision, and 4-ms inference latency, with a memory footprint of 19 MB, outperforming static AI firewalls in both efficiency and scalability. The proposed NGFW framework offers a resilient and scalable solution for real-time attack detection in resource-constrained environments without requiring frequent model retraining. Future enhancements include hybrid shallow–deep model architectures, continuous federated learning for decentralized adaptability, and the integration of explainable AI to enhance transparency and trustworthiness in edge security deployments.
Embargo
Air pollution, socioeconomic development, and ischaemic heart disease mortality in Asia
(Springer Science and Business Media, 2026-07) Cooray, D; Jayathilaka, R
Ischaemic Heart Disease (IHD) remains one of the leading causes of global mortality. Recent evidence indicates a persistent upward trend in IHD mortality in Asia, despite declines in many other regions. Increasing exposure to ambient air pollution, together with socioeconomic and behavioural transitions, suggests that environmental and social determinants may play an important role in shaping regional cardiovascular outcomes. This study investigates the combined influence of air pollution, socioeconomic development, behavioural factors, and pathological conditions on IHD mortality across 31 Asian countries from 2000 to 2021. Panel regression techniques were employed to capture temporal dynamics and cross-country heterogeneity. These were complemented by country-specific linear regressions to identify differential national-level effects. The analysis focuses on ambient particulate matter (PM2.5), alcohol consumption, anaemia prevalence, and the Human Development Index (HDI). The results show that a one-litre increase in per-capita alcohol consumption is associated with approximately eight additional IHD deaths per 100,000 population. Higher anaemia prevalence also significantly increases mortality risk. In addition, a 1% increase in PM2.5 exposure associated with approximately 0.4 additional IHD deaths per 100,000 population, highlighting the cardiovascular burden of poor air quality. In contrast, higher levels of socioeconomic development, reflected by improvements in HDI, are associated with a substantial reduction in IHD mortality. These findings highlight the need for integrated policy responses that combine air quality regulation with health, nutrition, and socioeconomic development strategies. Strengthening air pollution control, together with broader socioeconomic interventions, is essential for reducing the growing cardiovascular burden in Asia.
Embargo
Alcohol Consumption and Stroke Mortality: Global Patterns, Risks and Public Health Implications
(Springer, 2025-05-07) Kolonne, T; Mudalige, K; Dissanayaka, G; Rathnayake, K; Jayathilaka, R; Rajamanthri, L; Wickramaarachchi, C
Globally, stroke remains a leading cause of mortality and disability, while alcohol consumption continues to vary widely across regions, prompting concern over its health impacts. This study examines the association between different alcoholic beverages and stroke mortality, using secondary data from 1990 to 2020. Alcohol consumption and stroke death rates across 189 countries were categorized into five levels, from very high to very low, and averaged over two periods (1990–1999 and 2011–2020). Multiple Correspondence Analysis (MCA) was applied to assess relationships among four categorical variables. The findings indicate a significant association between very high alcohol consumption and increased stroke mortality, with eight countries showing elevated death rates. Conversely, moderate beer consumption was linked to reduced stroke mortality, suggesting nuanced effects based on beverage type and quantity. These insights offer a foundation for targeted public health policies and emphasize the need for further investigation into the mechanisms driving alcohol-related stroke risks.
Open Access
Ambient Backscatter- and Simultaneous Wireless Information and Power Transfer-Enabled Switch for Indoor Internet of Things Systems
(Multidisciplinary Digital Publishing Institute (MDPI), 2025-01-06) Sooriarachchi, V.P; Perera, T.D.P; Jayakody, D.N.K
Indoor Internet of Things (IoT) is considered as a crucial component of Industry 4.0, enabling devices and machine to communicate and share sensed data leading to increased efficiency, productivity, and automation. Increased energy efficiency is a significant focus within Industry 4.0, as it offers numerous benefits. To support this focus, we developed a hybrid switching mechanism to switch between energy harvesting techniques, ambient backscattering and Simultaneous Wireless Information and Power Transfer (SWIPT), which can be utilized within cooperative communications. To implement the proposed switching mechanism, we consider an indoor warehouse environment, where the moving sensor node transmits sensed data to the fixed relay located on the roof, which is then transmitted to an IoT gateway. The relay is equipped with the proposed switch to energize its communication capabilities while maintaining the expected quality of service at the IoT gateway. Simulation results illustrate the improved energy efficiency within the indoor communication setup while maintaining QoS at varying signal-to-noise (SNR) conditions.
Open Access
Anthocyanin (ATH)-incorporating polyvinylpyrrolidone-ethyl cellulose-(2-hydroxypropyl)-β-cyclodextrin (PVP–EC–BCD) nanofiber-based pH sensor for ocular pH detection during accidental chemical spills
(Royal Society of Chemistry, 2026-02-03) Sandaruwan, B; Liyanage, R; Costha, P; Dassanayake, Rohan S; Wijesinghe, R, E; Herath H.M.L.P.B; Nalin de Silva K.M.; de Silva, Rohini M; Rajapaksha, Suranga M; Wijenayake, U; Manatunga, Danushika C
The existing ocular pH detection methods encounter numerous limitations, including low accuracy, poor sensitivity across a wide pH range, and patient discomfort, highlighting the need for innovative approaches. A novel biosensor for ocular pH detection has been developed to assess ocular health and chemical injuries in clinical settings. This study uses the pH-sensitive properties of anthocyanins (ATHs), natural pigments extracted from butterfly pea flowers, to develop a novel pH-responsive nanofiber mat. ATHs are integrated into a polymer blend containing polyvinylpyrrolidone (PVP), ethyl cellulose (EC), and (2-hydroxypropyl)-β-cyclodextrin (BCD) to fabricate electrospun nanofibers. The acquired characterization, employing scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), and thermogravimetric analysis (TGA), confirmed the successful fabrication of the ATH-infused nanofibers with a mean diameter ranging from 121 to 396 nm. Four formulations were tested: PVP:EC:BCD:ATH (18 ppm), PVP:EC:BCD:ATH (25 ppm), PVP:EC:BCD:ATH (35 ppm), and PVP:EC:BCD:ATH (50 ppm). Among them, the 50 ppm ATH-incorporating nanofiber mat exhibited the best performance in terms of color clarity, response time, and pH sensitivity. The fabricated 50 ppm ATH incorporating nanofiber mat demonstrated a rapid pH response time of less than 5 seconds (s) while exhibiting a color variation from pink to blue to green across the pH range of 1 to 12, providing a rapid and accurate method for visual pH detection. Based on the color performance of the 50 ppm ATH-incorporating system, a standardized color reference chart was developed to serve as a practical and visual guide for estimating pH levels in clinical applications. Zebrafish toxicity assays were conducted further to validate the safety and biocompatibility of the developed sensor, revealing no significant toxic effects across the range of ATH concentrations.
Open Access
Assessing the influence of diverse skills on employability outcomes for IT undergraduates
(Public Library of Science, 2026-04-10) Senadheera, D; Wisenthige, K
Rapid technological advancements have reshaped the global job market, emphasizing the importance of specialized competencies such as user interface (UI) and user experience (UX) design, alongside technical and interpersonal skills.. This study examines how UI/UX skills (UIUX), soft skills (SS), and technical skills (TS) influence the employability (EP) of IT undergraduates in Sri Lanka, addressing a notable gap in existing literature that often examines these competencies in isolation and predominantly within Western contexts. The current study offers a detailed examination of employability determinants in Sri Lanka’s IT sector by incorporating gender as a moderating factor and investigating the mediating roles of self-efficacy (SE) and proficiency levels (LP). The collection of data involved 345 IT undergraduates participating in structured surveys, which were subsequently analyzed using partial least squares structural equation modelling (PLS-SEM). The results demonstrate that gender significantly affects the relationship between soft skills and technical skills with employability, underscoring differences in the assessment of these competencies among different genders. Moreover, the degree of proficiency influences the connection between technical skills and employability, yet it does not play a significant mediating role in the relationship between soft skills and UI/UX employability. Self-efficacy has proven to be a significant mediator across various skill categories UI/UX, soft, and technical highlighting its essential function in converting competencies into career success. This work seeks to add to existing knowledge by tackling the main significant gap of examining the combined effect of UI/UX, soft, and technical skills on employability. This study contributes to the theoretical understanding of employability by presenting an integrated model that elucidates the complex interactions among skills, mediators, and gender within the Sri Lankan IT sector. The results provide actionable insights for educators, policymakers, and industry leaders, advocating for curriculum alignment with industry needs and the promotion of self-efficacy through mentorship and experiential learning.
Open Access
Automated design of reinforced concrete dapped-end connections using hybrid deep learning and generative AI augmentation
(Elsevier Ltd, 2026-04-15) Dharmawansha, S; Herath, S; Fernando P.L.N; Meddage D.P.P.; Rajapakse, C
Dapped-end connections, also known as half-joints or Gerber beams, are widely used yet structurally vulnerable elements in precast concrete structures due to high stress concentrations near the re-entrant corner. Therefore, a comprehensive assessment of the load-bearing capacity of dapped-end connections is important to ensure structural integrity and mitigate the risk of failure. Although prior studies have explored their behaviour through analytical and experimental methods, the application of data-driven approaches remains limited due to the availability of limited experimental data, which constrains the predictive accuracy and generalisation of Machine Learning (ML) models. This study presents a novel approach that integrates numerical simulation with Conditional Tabular Generative Adversarial Network (CTGAN)-based data augmentation to enhance prediction accuracy and model generalisation. A numerical database containing 720 results was developed, which was expanded with 680 augmented data using CTGAN. The combined dataset of 1400 instances was used to train Artificial Neural Network (ANN), Genetic Algorithm-ANN (GA-ANN), and Particle Swarm Optimisation-ANN (PSO-ANN) models. The hybrid models outperformed the standalone ANN, with GA-ANN achieving the highest accuracy (testing R2 = 0.961). The trained models were separately validated using 64 unseen experimental datasets, which shows the improved generalisation of the models through augmentation. Shapley Additive Explanations analysis reveals that the GA-ANN model predictions aligned with the principles underlying the compatibility of deformations of dapped-ends. Further, a novel ML-assisted design model was developed, which predicts multiple solutions for a given design problem, assisting in the optimisation of connection design.
Embargo
Balancing Maintenance Costs and Degradation Levels in Building Elements
(American Society of Civil Engineers (ASCE), 2025-10-06) Dissanayake, D; Dias W.P.S
Maintenance strategies require a balance between maintaining assets at acceptable levels of service and minimizing maintenance costs. The latter is imperative especially for building elements, since they require less maintenance than machinery, and carrying out such maintenance can be disruptive and costly. This paper reports an inquiry into optimal maintenance strategies for 12 building elements, ranging from structural through finishes and opening to services ones, each with a differing service life. Element degradation was predicted using Markov chain models, while repair costs were based on lost value ratios associated with five discrete condition ratings, separately defined for each of the elements. The strategies considered were based on three different criteria, namely, minimizing maintenance cost, achieving the best the overall condition rating at end of service life, and optimizing a rating that combined both cost and quality. The maintenance cycles considered were 5, 10, and 20 years, and also the zero-maintenance case, costed at both zero and 5% net discount rates. The repair combinations involved upgrading elements in a single given Condition 2, 3, 4, or 5 to the as new Condition 1 or upgrading elements in all such conditions to Condition 1. It was found that performing regular maintenance on concrete slabs, floor tiles, timber windows, and especially ceiling fans and fan regulators was in fact cheaper than not maintaining them (and having to renovate at the end of service life), while a 20 year cycle was found to be optimal for the other elements studied.
Embargo
Bi-directional long short-term memory based ensemble deep learning framework for non-linear steam turbine power forecasting: a biomass fuelled case study
(Elsevier Ltd, 2026-04-10) Perera, H; Jayasekara, S; Wijesinghe, R.E; Silva, B. N; Cha, H
In palm oil manufacturing, steam turbines powered by biomass fuel are central to energy generation. However, fluctuating load demands and temporal variations lead to inefficiencies, while limited and variable supply of biomass waste constrains boiler feed flexibility. Current index-based boiler feeding methods overlook actual load demands and waste availability, resulting in significant energy wastage. This study presents a novel ensemble deep learning model combining Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Units (GRU) with Attention Layers, trained on an eight-year operational dataset with structured preprocessing and feature selection, to forecast steam turbine power generation. The model captures complex non-linear temporal patterns more effectively than conventional and standalone ML models, achieving a Root Mean Square Error (RMSE) of 0.0684, Mean Absolute Error (MAE) of 0.0414, and an R-squared (R2) value of 0.9832, which outperformed eight benchmark models by approximately 25% in prediction accuracy. Additionally, the framework incorporates operational parameters such as kVA, total energy, and Fresh Fruit Bunch (FFB) production to dynamically optimise biomass feed rates, balancing energy output with resource availability. This approach minimises energy wastage, reduces grid reliance, and promotes both sustainability and profitability.